Why brands aren't scared of AI slop anymore
For a while, brands got caught using AI and rushed to apologize. Lately, some of the biggest names have posted obviously AI-made ads with a shrug and a disclaimer in the corner. The tone has shifted from "whoops" to "yeah, we used AI-and?"
What changed
The shame wore off. People have seen so much shiny, uncanny imagery that the AI look feels common. Inside big companies, teams are desensitized to backlash cycles that burn hot for 48 hours, then vanish.
Innovation theater is driving decisions. Execs want "AI" on slide decks. Agencies pitch AI-first to look modern. Teams chase OKRs tied to speed and cost, not craft. "We shipped fast" beats "we made it great" in too many rooms.
The math favors shortcuts. AI comps reduce shoot days, talent fees, and post. Even if the ad looks worse, the cost delta is big-and most outrage doesn't dent sales meaningfully. Until that changes, some brands will tolerate slop.
Platforms reward volume. Feeds are relentless. If a post lives for six days, speed wins the slot. Quality is a brand asset; reach is a weekly quota. That tension is now visible in your timeline.
But audiences still notice
Consumers keep calling out "AI tells"-warped hands, broken type, off lighting, wrong labels. Skepticism around obvious AI in ads is real and persistent. Surveys show people remain more concerned than excited about AI in daily life, which tracks with the tone of public comments on these campaigns. Pew Research has documented this cautious sentiment.
What this means for creatives
You're negotiating two timelines: short-term content quotas and long-term brand equity. Your job isn't to reject AI blindly-it's to protect the craft where it matters and use the tech where it actually helps.
Set clear guardrails (before the brief lands)
- Define "hero" vs. "supporting" assets. Hero: no raw AI imagery in final. Supporting: AI allowed with strict QA and human polish.
- Codify non-negotiables. Product accuracy, legal marks, brand codes (color, typeface, grid), and anatomy fidelity are zero-tolerance zones.
- Version control. Log models, prompts, seeds, and edits. You'll need provenance if a claim or dispute arises. The FTC's guidance is clear: be truthful and substantiated.
Pre-flight QA: kill the tells
- Hands, ears, eyes. Count digits, look for jewelry melt, check reflections and lens distortions.
- Typography. Kerning, repeated letters, warped glyphs, inconsistent baselines.
- Packaging and product. Label copy, nutrition/legal icons, serials, safety marks, skewed logos.
- Physics and lighting. Shadow direction, specular highlights, depth of field, fabric folds.
- Continuity. Props, wardrobe, hair length, brand colors across frames.
- Claims and compliance. Any text in-frame must match approved language.
Where AI actually works (without wrecking the craft)
- Concepting. Moodboards, visual territories, quick storyboards, product styling ideas.
- Versioning. Background swaps, minor set extensions, social cutdowns, copy variants for testing.
- Previs and prototyping. Rough comps to sell-in an idea before booking a studio or building a set.
- 3D hybrids. Real product renders with consistent lighting, then subtle AI cleanup-not the other way around.
Have the "speed vs. brand" talk with stakeholders
Use a simple script: "We'll hit the deadline by using AI for exploration and volume. For hero placements, we keep human craft to protect distinctiveness and avoid brand risk. Here's the cost if we cut corners-and the lift we've seen when we don't."
Measure what matters
- Brand lift and favorability in paid placements (not just CTR).
- Distinctiveness (recognition of brand codes in forced-exposure tests).
- Error rate (compliance or product inaccuracies per campaign).
- Cycle time saved by AI on non-hero assets-document the hours.
If you must disclose AI, do it with intent
Don't bury it in tiny print that reads like guilt. Own the choice and explain the value to the viewer when relevant (speed to inform, not an excuse for low effort). If the result looks cheap, the disclosure won't save it-fix the work.
A realistic production model
- Week 1: Strategy, brand codes audit, AI concepting, select territories.
- Week 2: Shoot or 3D for hero assets; AI for support variants and situational tests.
- Week 3: Human polish, compliance checks, final QA, then scale versions with controlled AI.
Risks to flag early
- Legal/provenance. Training data disputes, likeness rights, unlicensed textures or fonts.
- Brand blur. Generic AI aesthetics erode distinctiveness over time.
- Crisis cost. Outrage fades, but screenshots live forever. Factor PR lift studies and escalation playbooks into planning.
Tools and training worth your time
- AI for Creatives - practical ways to integrate AI into design workflows without sacrificing brand craft.
- AI for PR & Communications - useful for shaping disclosure policies and prepping for backlash.
The bottom line
Some brands will keep posting AI slop because the spreadsheet says it's fine. That doesn't have to be your work. Use AI where it speeds clarity and options, then protect the moments that define the brand. Speed is a tactic. Craft is the moat.
Your membership also unlocks:
